sysusers: add another column to sysusers files for the home directory

5 files changed, 468 insertions, 67 deletions

diff --git a/man/sysusers.d.xml b/man/sysusers.d.xml
index 58f24a62f5..1832ecf0e6 100644
--- a/man/sysusers.d.xml
+++ b/man/sysusers.d.xml
@@ -77,13 +77,14 @@
                 configuration.</para>
 
                 <para>The file format is one line per user or group
-                containing name, ID and GECOS field description:</para>
+                containing name, ID, GECOS field description and home directory:</para>
 
                 <programlisting># Type Name ID GECOS
 u httpd 440 "HTTP User"
 u authd /usr/bin/authd "Authorization user"
 g input - -
-m authd input</programlisting>
+m authd input
+u root 0 "Superuser" /root</programlisting>
 
                 <refsect2>
                         <title>Type</title>
@@ -103,11 +104,13 @@ m authd input</programlisting>
                                         bearing the same name. The
                                         user's shell will be set to
                                         <filename>/sbin/nologin</filename>,
-                                        the home directory to
-                                        <filename>/</filename>. The
-                                        account will be created
-                                        disabled, so that logins are
-                                        not allowed.</para></listitem>
+                                        the home directory to the
+                                        specified home directory, or
+                                        <filename>/</filename> if none
+                                        is given. The account will be
+                                        created disabled, so that
+                                        logins are not
+                                        allowed.</para></listitem>
                                 </varlistentry>
 
                                 <varlistentry>
@@ -160,8 +163,8 @@ m authd input</programlisting>
                         <varname>g</varname> the numeric 32bit UID or
                         GID of the user/group. Do not use IDs 65535 or
                         4294967295, as they have special placeholder
-                        meanings. Specify "-" for automatic UID/GID
-                        allocation for the user or
+                        meanings. Specify <literal>-</literal> for
+                        automatic UID/GID allocation for the user or
                         group. Alternatively, specify an absolute path
                         in the file system. In this case the UID/GID
                         is read from the path's owner/group. This is
@@ -183,7 +186,24 @@ m authd input</programlisting>
 
                         <para>Only applies to lines of type
                         <varname>u</varname> and should otherwise be
-                        left unset.</para>
+                        left unset, or be set to
+                        <literal>-</literal>.</para>
+                </refsect2>
+
+                <refsect2>
+                        <title>Home Directory</title>
+
+                        <para>The home directory for a new system
+                        user. If omitted defaults to the root
+                        directory. It is recommended to not
+                        unnecessarily specify home directories for
+                        system users, unless software strictly
+                        requires one to be set.</para>
+
+                        <para>Only applies to lines of type
+                        <varname>u</varname> and should otherwise be
+                        left unset, or be set to
+                        <literal>-</literal>.</para>
                 </refsect2>
 
         </refsect1>
diff --git a/src/shared/util.c b/src/shared/util.c
index 18d40f398f..2bb3b5ef1c 100644
--- a/src/shared/util.c
+++ b/src/shared/util.c
@@ -6934,3 +6934,208 @@ int is_symlink(const char *path) {
 
         return 0;
 }
+
+int unquote_first_word(const char **p, char **ret) {
+        _cleanup_free_ char *s = NULL;
+        size_t allocated = 0, sz = 0;
+
+        enum {
+                START,
+                VALUE,
+                VALUE_ESCAPE,
+                SINGLE_QUOTE,
+                SINGLE_QUOTE_ESCAPE,
+                DOUBLE_QUOTE,
+                DOUBLE_QUOTE_ESCAPE,
+                SPACE,
+        } state = START;
+
+        assert(p);
+        assert(*p);
+        assert(ret);
+
+        /* Parses the first word of a string, and returns it in
+         * *ret. Removes all quotes in the process. When parsing fails
+         * (because of an uneven number of quotes or similar), leaves
+         * the pointer *p at the first invalid character. */
+
+        for (;;) {
+                char c = **p;
+
+                switch (state) {
+
+                case START:
+                        if (c == 0)
+                                goto finish;
+                        else if (strchr(WHITESPACE, c))
+                                break;
+
+                        state = VALUE;
+                        /* fallthrough */
+
+                case VALUE:
+                        if (c == 0)
+                                goto finish;
+                        else if (c == '\'')
+                                state = SINGLE_QUOTE;
+                        else if (c == '\\')
+                                state = VALUE_ESCAPE;
+                        else if (c == '\"')
+                                state = DOUBLE_QUOTE;
+                        else if (strchr(WHITESPACE, c))
+                                state = SPACE;
+                        else {
+                                if (!GREEDY_REALLOC(s, allocated, sz+2))
+                                        return -ENOMEM;
+
+                                s[sz++] = c;
+                        }
+
+                        break;
+
+                case VALUE_ESCAPE:
+                        if (c == 0)
+                                return -EINVAL;
+
+                        if (!GREEDY_REALLOC(s, allocated, sz+2))
+                                return -ENOMEM;
+
+                        s[sz++] = c;
+                        state = VALUE;
+
+                        break;
+
+                case SINGLE_QUOTE:
+                        if (c == 0)
+                                return -EINVAL;
+                        else if (c == '\'')
+                                state = VALUE;
+                        else if (c == '\\')
+                                state = SINGLE_QUOTE_ESCAPE;
+                        else {
+                                if (!GREEDY_REALLOC(s, allocated, sz+2))
+                                        return -ENOMEM;
+
+                                s[sz++] = c;
+                        }
+
+                        break;
+
+                case SINGLE_QUOTE_ESCAPE:
+                        if (c == 0)
+                                return -EINVAL;
+
+                        if (!GREEDY_REALLOC(s, allocated, sz+2))
+                                return -ENOMEM;
+
+                        s[sz++] = c;
+                        state = SINGLE_QUOTE;
+                        break;
+
+                case DOUBLE_QUOTE:
+                        if (c == 0)
+                                return -EINVAL;
+                        else if (c == '\"')
+                                state = VALUE;
+                        else if (c == '\\')
+                                state = DOUBLE_QUOTE_ESCAPE;
+                        else {
+                                if (!GREEDY_REALLOC(s, allocated, sz+2))
+                                        return -ENOMEM;
+
+                                s[sz++] = c;
+                        }
+
+                        break;
+
+                case DOUBLE_QUOTE_ESCAPE:
+                        if (c == 0)
+                                return -EINVAL;
+
+                        if (!GREEDY_REALLOC(s, allocated, sz+2))
+                                return -ENOMEM;
+
+                        s[sz++] = c;
+                        state = DOUBLE_QUOTE;
+                        break;
+
+                case SPACE:
+                        if (c == 0)
+                                goto finish;
+                        if (!strchr(WHITESPACE, c))
+                                goto finish;
+
+                        break;
+                }
+
+                (*p) ++;
+        }
+
+finish:
+        if (!s) {
+                *ret = NULL;
+                return 0;
+        }
+
+        s[sz] = 0;
+        *ret = s;
+        s = NULL;
+
+        return 1;
+}
+
+int unquote_many_words(const char **p, ...) {
+        va_list ap;
+        char **l;
+        int n = 0, i, c, r;
+
+        /* Parses a number of words from a string, stripping any
+         * quotes if necessary. */
+
+        assert(p);
+
+        /* Count how many words are expected */
+        va_start(ap, p);
+        for (;;) {
+                if (!va_arg(ap, char **))
+                        break;
+                n++;
+        }
+        va_end(ap);
+
+        if (n <= 0)
+                return 0;
+
+        /* Read all words into a temporary array */
+        l = newa0(char*, n);
+        for (c = 0; c < n; c++) {
+
+                r = unquote_first_word(p, &l[c]);
+                if (r < 0) {
+                        int j;
+
+                        for (j = 0; j < c; j++) {
+                                free(l[j]);
+                                return r;
+                        }
+                }
+
+                if (r == 0)
+                        break;
+        }
+
+        /* If we managed to parse all words, return them in the passed
+         * in parameters */
+        va_start(ap, p);
+        for (i = 0; i < n; i++) {
+                char **v;
+
+                v = va_arg(ap, char **);
+                assert(v);
+
+                *v = l[i];
+        }
+        va_end(ap);
+
+        return c;
+}
diff --git a/src/shared/util.h b/src/shared/util.h
index bd8bbb268f..43f4b089b4 100644
--- a/src/shared/util.h
+++ b/src/shared/util.h
@@ -128,6 +128,8 @@ bool streq_ptr(const char *a, const char *b) _pure_;
 
 #define newa(t, n) ((t*) alloca(sizeof(t)*(n)))
 
+#define newa0(t, n) ((t*) alloca0(sizeof(t)*(n)))
+
 #define newdup(t, p, n) ((t*) memdup_multiply(p, sizeof(t), (n)))
 
 #define malloc0(n) (calloc((n), 1))
@@ -967,4 +969,7 @@ bool is_localhost(const char *hostname);
 
 int take_password_lock(const char *root);
 
-int is_symlink(const char *path);
\ No newline at end of file
+int is_symlink(const char *path);
+
+int unquote_first_word(const char **p, char **ret);
+int unquote_many_words(const char **p, ...) _sentinel_;
diff --git a/src/sysusers/sysusers.c b/src/sysusers/sysusers.c
index f78fb4fff3..eedc1e067e 100644
--- a/src/sysusers/sysusers.c
+++ b/src/sysusers/sysusers.c
@@ -51,6 +51,7 @@ typedef struct Item {
         char *uid_path;
         char *gid_path;
         char *description;
+        char *home;
 
         gid_t gid;
         uid_t uid;
@@ -556,19 +557,19 @@ static int write_files(void) {
                                 .pw_uid = i->uid,
                                 .pw_gid = i->gid,
                                 .pw_gecos = i->description,
+
+                                /* "x" means the password is stored in
+                                 * the shadow file */
                                 .pw_passwd = (char*) "x",
-                        };
 
-                        /* Initialize the home directory and the shell
-                         * to nologin, with one exception: for root we
-                         * patch in something special */
-                        if (i->uid == 0) {
-                                n.pw_shell = (char*) "/bin/sh";
-                                n.pw_dir = (char*) "/root";
-                        } else {
-                                n.pw_shell = (char*) "/sbin/nologin";
-                                n.pw_dir = (char*) "/";
-                        }
+                                /* We default to the root directory as home */
+                                .pw_dir = i->home ? i->home : (char*) "/",
+
+                                /* Initialize the shell to nologin,
+                                 * with one exception: for root we
+                                 * patch in something special */
+                                .pw_shell = i->uid == 0 ? (char*) "/bin/sh" : (char*) "/sbin/nologin",
+                        };
 
                         errno = 0;
                         if (putpwent(&n, passwd) != 0) {
@@ -1292,6 +1293,9 @@ static bool item_equal(Item *a, Item *b) {
         if (a->gid_set && a->gid != b->gid)
                 return false;
 
+        if (!streq_ptr(a->home, b->home))
+                return false;
+
         return true;
 }
 
@@ -1330,6 +1334,9 @@ static bool valid_user_group_name(const char *u) {
 
 static bool valid_gecos(const char *d) {
 
+        if (!d)
+                return false;
+
         if (!utf8_is_valid(d))
                 return false;
 
@@ -1343,6 +1350,30 @@ static bool valid_gecos(const char *d) {
         return true;
 }
 
+static bool valid_home(const char *p) {
+
+        if (isempty(p))
+                return false;
+
+        if (!utf8_is_valid(p))
+                return false;
+
+        if (string_has_cc(p, NULL))
+                return false;
+
+        if (!path_is_absolute(p))
+                return false;
+
+        if (!path_is_safe(p))
+                return false;
+
+        /* Colons are used as field separators, and hence not OK */
+        if (strchr(p, ':'))
+                return false;
+
+        return true;
+}
+
 static int parse_line(const char *fname, unsigned line, const char *buffer) {
 
         static const Specifier specifier_table[] = {
@@ -1353,27 +1384,34 @@ static int parse_line(const char *fname, unsigned line, const char *buffer) {
                 {}
         };
 
-        _cleanup_free_ char *action = NULL, *name = NULL, *id = NULL, *resolved_name = NULL;
+        _cleanup_free_ char *action = NULL, *name = NULL, *id = NULL, *resolved_name = NULL, *description = NULL, *home = NULL;
         _cleanup_(item_freep) Item *i = NULL;
         Item *existing;
         Hashmap *h;
-        int r, n = -1;
+        int r;
+        const char *p;
 
         assert(fname);
         assert(line >= 1);
         assert(buffer);
 
-        r = sscanf(buffer,
-                   "%ms %ms %ms %n",
-                   &action,
-                   &name,
-                   &id,
-                   &n);
-        if (r < 2) {
+        /* Parse columns */
+        p = buffer;
+        r = unquote_many_words(&p, &action, &name, &id, &description, &home, NULL);
+        if (r < 0) {
                 log_error("[%s:%u] Syntax error.", fname, line);
-                return -EIO;
+                return r;
+        }
+        if (r < 2) {
+                log_error("[%s:%u] Missing action and name columns.", fname, line);
+                return -EINVAL;
+        }
+        if (*p != 0) {
+                log_error("[%s:%u] Trailing garbage.", fname, line);
+                return -EINVAL;
         }
 
+        /* Verify action */
         if (strlen(action) != 1) {
                 log_error("[%s:%u] Unknown modifier '%s'", fname, line, action);
                 return -EINVAL;
@@ -1384,6 +1422,7 @@ static int parse_line(const char *fname, unsigned line, const char *buffer) {
                 return -EBADMSG;
         }
 
+        /* Verify name */
         r = specifier_printf(name, specifier_table, NULL, &resolved_name);
         if (r < 0) {
                 log_error("[%s:%u] Failed to replace specifiers: %s", fname, line, name);
@@ -1395,11 +1434,20 @@ static int parse_line(const char *fname, unsigned line, const char *buffer) {
                 return -EINVAL;
         }
 
-        if (n >= 0) {
-                n += strspn(buffer+n, WHITESPACE);
+        /* Simplify remaining columns */
+        if (isempty(id) || streq(id, "-")) {
+                free(id);
+                id = NULL;
+        }
+
+        if (isempty(description) || streq(description, "-")) {
+                free(description);
+                description = NULL;
+        }
 
-                if (STR_IN_SET(buffer + n, "", "-"))
-                        n = -1;
+        if (isempty(home) || streq(home, "-")) {
+                free(home);
+                home = NULL;
         }
 
         switch (action[0]) {
@@ -1408,13 +1456,19 @@ static int parse_line(const char *fname, unsigned line, const char *buffer) {
                 _cleanup_free_ char *resolved_id = NULL;
                 char **l;
 
-                r = hashmap_ensure_allocated(&members, string_hash_func, string_compare_func);
-                if (r < 0)
-                        return log_oom();
-
                 /* Try to extend an existing member or group item */
 
-                if (!id || streq(id, "-")) {
+                if (description) {
+                        log_error("[%s:%u] Lines of type 'm' don't take a GECOS field.", fname, line);
+                        return -EINVAL;
+                }
+
+                if (home) {
+                        log_error("[%s:%u] Lines of type 'm' don't take a home directory field.", fname, line);
+                        return -EINVAL;
+                }
+
+                if (!id) {
                         log_error("[%s:%u] Lines of type 'm' require a group name in the third field.", fname, line);
                         return -EINVAL;
                 }
@@ -1430,10 +1484,9 @@ static int parse_line(const char *fname, unsigned line, const char *buffer) {
                         return -EINVAL;
                 }
 
-                if (n >= 0) {
-                        log_error("[%s:%u] Lines of type 'm' don't take a GECOS field.", fname, line);
-                        return -EINVAL;
-                }
+                r = hashmap_ensure_allocated(&members, string_hash_func, string_compare_func);
+                if (r < 0)
+                        return log_oom();
 
                 l = hashmap_get(members, resolved_id);
                 if (l) {
@@ -1476,15 +1529,12 @@ static int parse_line(const char *fname, unsigned line, const char *buffer) {
                 if (!i)
                         return log_oom();
 
-                if (id && !streq(id, "-")) {
-
+                if (id) {
                         if (path_is_absolute(id)) {
-                                i->uid_path = strdup(id);
-                                if (!i->uid_path)
-                                        return log_oom();
+                                i->uid_path = id;
+                                id = NULL;
 
                                 path_kill_slashes(i->uid_path);
-
                         } else {
                                 r = parse_uid(id, &i->uid);
                                 if (r < 0) {
@@ -1496,40 +1546,53 @@ static int parse_line(const char *fname, unsigned line, const char *buffer) {
                         }
                 }
 
-                if (n >= 0) {
-                        i->description = unquote(buffer+n, "\"");
-                        if (!i->description)
-                                return log_oom();
+                if (description) {
+                        if (!valid_gecos(description)) {
+                                log_error("[%s:%u] '%s' is not a valid GECOS field.", fname, line, description);
+                                return -EINVAL;
+                        }
+
+                        i->description = description;
+                        description = NULL;
+                }
 
-                        if (!valid_gecos(i->description)) {
-                                log_error("[%s:%u] '%s' is not a valid GECOS field.", fname, line, i->description);
+                if (home) {
+                        if (!valid_home(home)) {
+                                log_error("[%s:%u] '%s' is not a valid home directory field.", fname, line, home);
                                 return -EINVAL;
                         }
+
+                        i->home = home;
+                        home = NULL;
                 }
 
                 h = users;
                 break;
 
         case ADD_GROUP:
-                r = hashmap_ensure_allocated(&groups, string_hash_func, string_compare_func);
-                if (r < 0)
-                        return log_oom();
 
-                if (n >= 0) {
+                if (description) {
                         log_error("[%s:%u] Lines of type 'g' don't take a GECOS field.", fname, line);
                         return -EINVAL;
                 }
 
+                if (home) {
+                        log_error("[%s:%u] Lines of type 'g' don't take a home directory field.", fname, line);
+                        return -EINVAL;
+                }
+
+                r = hashmap_ensure_allocated(&groups, string_hash_func, string_compare_func);
+                if (r < 0)
+                        return log_oom();
+
                 i = new0(Item, 1);
                 if (!i)
                         return log_oom();
 
-                if (id && !streq(id, "-")) {
-
+                if (id) {
                         if (path_is_absolute(id)) {
-                                i->gid_path = strdup(id);
-                                if (!i->gid_path)
-                                        return log_oom();
+                                i->gid_path = id;
+                                id = NULL;
 
                                 path_kill_slashes(i->gid_path);
                         } else {
@@ -1543,7 +1606,6 @@ static int parse_line(const char *fname, unsigned line, const char *buffer) {
                         }
                 }
 
-
                 h = groups;
                 break;
         default:
diff --git a/src/test/test-util.c b/src/test/test-util.c
index e876248935..b74abac154 100644
--- a/src/test/test-util.c
+++ b/src/test/test-util.c
@@ -1101,6 +1101,113 @@ static void test_execute_directory(void) {
         rm_rf_dangerous(tempdir, false, true, false);
 }
 
+static void test_unquote_first_word(void) {
+        const char *p, *original;
+        char *t;
+
+        p = original = "foobar waldo";
+        assert_se(unquote_first_word(&p, &t) > 0);
+        assert_se(streq(t, "foobar"));
+        free(t);
+        assert_se(p == original + 7);
+
+        assert_se(unquote_first_word(&p, &t) > 0);
+        assert_se(streq(t, "waldo"));
+        free(t);
+        assert_se(p == original + 12);
+
+        assert_se(unquote_first_word(&p, &t) == 0);
+        assert_se(!t);
+        assert_se(p == original + 12);
+
+        p = original = "\"foobar\" \'waldo\'";
+        assert_se(unquote_first_word(&p, &t) > 0);
+        assert_se(streq(t, "foobar"));
+        free(t);
+        assert_se(p == original + 9);
+
+        assert_se(unquote_first_word(&p, &t) > 0);
+        assert_se(streq(t, "waldo"));
+        free(t);
+        assert_se(p == original + 16);
+
+        assert_se(unquote_first_word(&p, &t) == 0);
+        assert_se(!t);
+        assert_se(p == original + 16);
+
+        p = original = "\"";
+        assert_se(unquote_first_word(&p, &t) == -EINVAL);
+        assert_se(p == original + 1);
+
+        p = original = "\'";
+        assert_se(unquote_first_word(&p, &t) == -EINVAL);
+        assert_se(p == original + 1);
+
+        p = original = "yay\'foo\'bar";
+        assert_se(unquote_first_word(&p, &t) > 0);
+        assert_se(streq(t, "yayfoobar"));
+        free(t);
+        assert_se(p == original + 11);
+
+        p = original = "   foobar   ";
+        assert_se(unquote_first_word(&p, &t) > 0);
+        assert_se(streq(t, "foobar"));
+        free(t);
+        assert_se(p == original + 12);
+}
+
+static void test_unquote_many_words(void) {
+        const char *p, *original;
+        char *a, *b, *c;
+
+        p = original = "foobar waldi piep";
+        assert_se(unquote_many_words(&p, &a, &b, &c, NULL) == 3);
+        assert_se(p == original + 17);
+        assert_se(streq_ptr(a, "foobar"));
+        assert_se(streq_ptr(b, "waldi"));
+        assert_se(streq_ptr(c, "piep"));
+        free(a);
+        free(b);
+        free(c);
+
+        p = original = "'foobar' wa\"ld\"i   ";
+        assert_se(unquote_many_words(&p, &a, &b, &c, NULL) == 2);
+        assert_se(p == original + 19);
+        assert_se(streq_ptr(a, "foobar"));
+        assert_se(streq_ptr(b, "waldi"));
+        assert_se(streq_ptr(c, NULL));
+        free(a);
+        free(b);
+
+        p = original = "";
+        assert_se(unquote_many_words(&p, &a, &b, &c, NULL) == 0);
+        assert_se(p == original);
+        assert_se(streq_ptr(a, NULL));
+        assert_se(streq_ptr(b, NULL));
+        assert_se(streq_ptr(c, NULL));
+
+        p = original = "  ";
+        assert_se(unquote_many_words(&p, &a, &b, &c, NULL) == 0);
+        assert_se(p == original+2);
+        assert_se(streq_ptr(a, NULL));
+        assert_se(streq_ptr(b, NULL));
+        assert_se(streq_ptr(c, NULL));
+
+        p = original = "foobar";
+        assert_se(unquote_many_words(&p, NULL) == 0);
+        assert_se(p == original);
+
+        p = original = "foobar waldi";
+        assert_se(unquote_many_words(&p, &a, NULL) == 1);
+        assert_se(p == original+7);
+        assert_se(streq_ptr(a, "foobar"));
+
+        p = original = "     foobar    ";
+        assert_se(unquote_many_words(&p, &a, NULL) == 1);
+        assert_se(p == original+15);
+        assert_se(streq_ptr(a, "foobar"));
+}
+
 int main(int argc, char *argv[]) {
         log_parse_environment();
         log_open();
@@ -1168,6 +1275,8 @@ int main(int argc, char *argv[]) {
         test_search_and_fopen_nulstr();
         test_glob_exists();
         test_execute_directory();
+        test_unquote_first_word();
+        test_unquote_many_words();
 
         return 0;
 }